The "AI writing taboo" wasn't a top down thing, the Chicago Manual of Style doesn't formally forbid it, it's more of a (true) grassroots thing, isn't it? I mean, I personally loath LLM-written text. If I begin to suspect that the putative author of an article had an LLM as load-bearing, I quit reading at the next sign of LLM-generation. I realize LLM text is as obviously AI written as it ever will be. I also hope that people who aren't billionaires or scammers will have their consciences nag them if they use LLMs to generate some crap.